RADI-06. SINGLE- VERSUS MULTI-FRACTION STEREOTACTIC RADIOSURGERY FOR BRAINSTEM METASTASES

BACKGROUND: For intracranial metastases with planning target volume (PTV) overlap of the brainstem (BSmet), the radiosurgical dose-fractionation that optimizes the therapeutic window is unknown. MATERIALS/METHODS: A retrospective review of brain metastases (BM) with/without BSmets treated with single-fraction stereotactic radiosurgery (SRS) or hypofractionated (2–5 fractions) radiosurgery (HF-SRS) between 2012–2016 was performed. Brainstem biologically effective doses (BED) and single-fraction equivalents of brainstem V10/V12 were calculated using α/β=3. Characteristics were compared between patients with/without BSmet and between SRS/HF-SRS cohorts using Wilcoxon rank sum, chi-square, or Fisher’s exact tests. Radiographic progression (RP) was assessed in patients with post-treatment contrasted MRI and defined as BSmet enlargement regardless of etiology (progression, radionecrosis, indeterminate). Kaplan-Meier estimates were compared between cohorts using log-rank test. RESULTS: 634 SRS/HF-SRS courses were identified, of which 59 (9.3%) treated ≥1 BSmet in 55 patients. BSmets occurred more commonly in patients with >4 BM (31% vs 10%, p< 0.001) and intracranial recurrence (39% vs 20%, p=0.003). BSmets were treated in 1 (22/59; 37%), 2 (1/59; 2%), or 5 (36/59; 61%) fractions. Age, KPS, and primary tumor site were balanced between SRS/HF-SRS cohorts. The HF-SRS cohort had significantly larger BSmet PTV (median 1.39cc vs 0.39cc, p=0.021), marginal dose (median 25Gy vs 15Gy, p< 0.001), brainstem V10 (median 1.60cc vs 0.47cc, p< 0.001), brainstem V12 (median 0.78cc vs 0.06cc, p< 0.001), and mean brainstem BED (median 9.27Gy3 vs 6.55Gy3, p=0.019). The SRS cohort was more likely to have prior whole brain radiotherapy (50% vs 14%, p=0.005) and restart steroids post-treatment (78% vs 41%, p=0.019). RP occurred in 6/17 vs 2/25 patients in the SRS vs HF-SRS cohorts, respectively (p=0.045). HF-SRS trended to higher freedom from RP (93% vs 74% @12mo; p=0.072). There was no overall survival difference (p=0.36). CONCLUSIONS: HF-SRS was associated with decreased RP and decreased likelihood of restarting steroids despite treating larger BSmets.